experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

[TF] Increase tolerance of elu tests on half precision numbers.

Browse Source 
Previously more precision was required than is supported by 16 bit floats.

PiperOrigin-RevId: 251612304
This commit is contained in:
A. Unique TensorFlower 2019-06-05 03:36:01 -07:00 committed by TensorFlower Gardener
parent 6db4b700fe
commit c47ad082f4
1 changed files with 2 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
tensorflow/python/kernel_tests/relu_op_test.py
View File

@ -437,7 +437,7 @@ class EluTest(test.TestCase):
def _testElu(self, np_features):
np_elu = self._npElu(np_features)
tf_elu = nn_ops.elu(np_features)
self.assertAllClose(np_elu, tf_elu)
self.assertAllCloseAccordingToType(np_elu, tf_elu)
self.assertShapeEqual(np_elu, tf_elu)
def testNumbersCPU(self):
@ -447,7 +447,6 @@ class EluTest(test.TestCase):
self._testElu(
np.array([[-9, 7, -5, 3, -1], [1, -3, 5, -7, 9]]).astype(t))
@test_util.disable_xla("Too much precision required for halfs")
def testNumbersGPU(self):
if not test.is_gpu_available():
self.skipTest("No GPU available")
@ -546,10 +545,9 @@ class SeluTest(test.TestCase):
def _testSelu(self, np_features):
np_selu = self._npSelu(np_features)
tf_selu = nn_ops.selu(np_features)
self.assertAllClose(np_selu, tf_selu)
self.assertAllCloseAccordingToType(np_selu, tf_selu)
self.assertShapeEqual(np_selu, tf_selu)
@test_util.disable_xla("Too much precision required for halfs")
def testNumbers(self):
for t in [np.float16, np.float32, np.float64]:
self._testSelu(